by Robert Varga | Leave us your feedback on this post!
PANTHEON.tech continued to be the leader in terms of contributions to the OpenDaylight codebase. While our focus remained on the core platform, we also dabbled into individual plugins to deliver significant performance, scalability and correctness improvements. We are therefore glad to be a significant part of the effort that went into the newest release of OpenDaylight – Sodium.
In Sodium, we have successfully transitioned OpenDaylight codebase to require Java 11 – an effort we have been spearheading since the mid-Neon timeframe. This allows our users to not only reap the runtime improvements of Java 11 (which was possible to do with Sodium), but it allows OpenDaylight code to take advantage of features available in Java 11.
Our continued stewardship of YANG Tools has seen us deliver multiple improvements and new features, most notable of which are:
YANG Parser: has been extended to support Errata 5617 leaf-ref path expressions, which violate RFC7950, but are used by various models seen in the wild — such as ETSI NFV models. YANG parser has also received some improvements in areas of CPU and memory usage when faced with large models, ranging from 10 to 25%, depending on the models and features used.
In-memory Data Tree: the technology underlying most datastore implementations. It has been improved in multiple areas, yielding an overall memory footprint reduction of 30%. This allows better out-of-the-box scalability.
Adoption of Java 11: Java has allowed PANTHEON.tech to deploy improvements to MD-SAL Binding runtime, resulting in measurably faster dispatch of methods across the system, providing benefits to all OpenDaylight plugin code.
We have also continued improvements in the Distributed Datastore. We achieved further improvements to the persistence format, reducing in-memory & on-disk footprint by as much as 25%.
Last but not least, we have taken a hard look at the OVSDB project and provided major refactors of the codebase. This has immensely improved its ability to scale the number of connected devices, as well as individual connection throughput.
If you would like a custom OpenDaylight integration, feel free to contact us!